How should evaluation be? Is a good evaluation of research also just? Towards the implementation of good evaluation